Hail risk is expected to increase in Central Europe. Findings on hail hazard changes often build on studying mesoscale environmental conditions or use area-integrated hail day statistics. Climatological insights into characteristics and trends on the storm-scale are still rare but can be highly risk-relevant. This study analyses 60,000 radar-based hailstorm tracks in the Swiss radar domain from 2002 to 2021, covering several European hail hotspots. We investigate whether long-term changes in hailstorm properties can be detected despite high natural variability and potential effects of radar modernisation. To do so, we caclulate and analyse hailstorm properties for the full sample and conditional on storm types. We explore distributional shifts in hailstorm properties using quantile regression and investigate whether statistical breakpoints overlap with instrumental changes. Our results indicate a consistent shift toward fewer but significantly larger and more organised hailstorms. We find no strong evidence of hailstorm properties systematically changing with radar modernisation, but show that very small, short-lived storms are less frequently detected in the later observation period. This shifts property distributions towards larger quantile values. To control for this, a complementary fixed-threshold hailstorm frequency analysis is done, revealing that the frequency of the most intense storms remained stable or even increased. Results further show that different hail-active years exhibit distinctive hailstorm characteristics, pinpointing research directions to further investigate the multi-scale processes of hailstorm property changes. These are needed to reduce the considerable uncertainties that still remain resulting from short data records, natural variability and technological constraints. Despite these limitations, our findings provide rare storm-scale evidence for a shift towards fewer, but larger, longer-lived, and more organised hailstorms in the Alpine region, resulting in increased damage potential.
Schulte-Fischedick et al. (Fri,) studied this question.